Allocation of multi-function resources within resource allocation and process control systems may be thought of as the management (i.e., administration, command, control, direction, governance, monitoring, regulation, etc.) of such multi-function resources (e.g., manufacturing tools, instruments, hardware, software, databases, communication/connectivity resources, transportation resources, facilities, utilities, inventories, etc.) among a variety of tasks within a process system.
Process systems may be arranged and implemented to manage large facilities, such as a manufacturing plant, a semiconductor fabrication facility, a mineral or crude oil refinery, or the like, as well as relatively smaller facilities, such as a corporate communications network, a data repository and management system, or the like. Such systems may be distributed or not, and typically include numerous modules tailored to manage various associated processes, wherein conventional means link these modules together to produce the distributed nature of the process system. This affords increased performance and a capability to expand or reduce the process system to satisfy changing needs.
Process systems are developed and tailored to satisfy wide ranges of process requirements, whether local, global or otherwise, and regardless of facility type. Such developers and users of such systems commonly have two principal objectives: (i) to centralize management/control of as many sub-processes or processes as possible to improve overall efficiency, and (ii) to support a common interface that communicates data among various modules managing/controlling or monitoring the processes, and also with any such centralized controller.
Each process, or group of associated sub-processes or processes, has certain input (e.g., data, diagnostics, feed, flow, power, etc.) and output (e.g., data, pressure, temperature, utilization parameters, etc.) characteristics associated therewith. These characteristics are measurable, and may be represented in a discernable manner.
Predictive control methodologies/techniques may be used to optimize certain processes as a function of such characteristics. Predictive control techniques may use algorithmic representations to estimate characteristic values (represented as parameters, variables, etc.) associated with them that can be used to better manage such process resources among a plurality of tasks.
Such optimization efforts only account mathematically for the tasks being performed and the process resources then used to resolve the same based upon statistical characteristics only, thereby failing to model and factor into the optimization effort both status and logistical data, as well as to account for human capabilities and interaction (i.e., functions, skills, qualifications, task preferences, track records and the like) that ultimately utilize the process resources to resolve the tasks. Conventional approaches can exhibit poor response to constantly changing or exigent circumstances, and as such fail to cooperatively optimize process resources, particularly process resources capable of performing multiple functions. What is needed in the art is a powerful and flexible means for dynamically analyzing and modifying process status in a real-time mode through allocation and reallocation of multifunction process resources among a plurality of tasks within a process system.
Using semiconductor fabrication as an example, in order to provide shortest cycle times, highest quality, timely-delivered cost-effective products that meet revenue growth plans, there is a continuous need to improve manufacturing processes and sub-processes, including the content and methods of delivering information to the operations staff.
Information about manufacturing tools and work in process (“WIP”) inventory are critical to the decision making process necessary to operate a semiconductor wafer manufacturing line. With complex multi-tool, multi-technology, multi-product resources (“multi-function resources”), a need exists in the industry for a system and method that allocate such multi-function resources among a plurality of tasks within fabrication facility so as to execute a flexible process or plan that responds to WIP mix, resource availability changes, associate work schedule, skill sets (e.g., “queue-jumping” hot lots, special work requests, etc.), and the like to meet the requirements of a “just-in-time” environment.
Stated more broadly, a measurement of process efficiency can be defined by how quickly demands by requesting tasks are satisfied through the allocation of process resources. Today, even though human operators assist in the allocation of resources to requesting tasks, decisions to allocate such resources are controlled by management (whether human management based upon periodic reports (e.g., daily, weekly, monthly or, even, quarterly), or automated management based upon periodic batched data, or some combination of the two) which reacts or decides based upon relatively stale data, rather than reacting/deciding dynamically.
For example, in a wetdeck process in semiconductor wafer fabrication it is economically advantageous to simultaneously process as many wafer lots as allowed by a wafer carrier, but not more than will fit into a furnace batch. Some types of wafer carriers have a carrier size of twenty four (24) while other types of wafer carrier have a carrier size of up to one hundred (100) wafers. A wafer carrier of one hundred (100) wafers could be loaded to complete the batch size requirement for up to four (4) different furnaces. The batch rules must be enforced to ensure than all wafers get a proper recipe. An error in batching will likely result in many wafers having to be scrapped.
It is also important that the re-clean time window requirements be observed when making up wafer batches. The importance of the re-clean time window requirements can be better understood by considering the following example. Assume that a first wafer lot and a second wafer lot both need the same wetdeck recipe. Also assume that an available wetdeck has a wetdeck carrier that is capable of holding all of the wafers from both wafer lots. Also assume that the first wafer lot and the second wafer lot are destined to go to different furnaces.
Processing both of the wafer lots simultaneously is a good idea as long as the re-clean time window associated with each wafer lot will not be exceeded. Re-clean time windows can be as short as one (1) hour. More typically, re-clean time windows can be as long as twenty four (24) hours. The variation in the length of re-clean time windows makes the efficient management of wetdeck resources even more difficult. In some cases, the opportunity for re-cleaning is not permitted at all. When re-cleaning is not permitted, the wafer lots will either make to the furnace in time or the wafer lots will be declared to be discrepant wafer lots. Depending upon the specific process involved, discrepant wafer lots may be required to be scrapped.
In any event it is very costly to re-clean material when the re-clean time window requirement is exceeded. The cost is due to the cost of chemicals, extra machine cycles, and the increased risk of processing errors and inappropriate batching. The successful operation of a wetdeck process in semiconductor wafer fabrication involves not just the type of materials used but also the timing of providing those materials.
Therefore, a need exists for a system and method for efficiently allocating multi-function resources for a system process in semiconductor wafer fabrication. A further need exists for a process system and related graphical user interface (“GUI”) through which management reacts timely relative to conventional systems based upon dynamic data. In particular, a need exists for GUIs, particularly, dispatch viewers, for association with systems for allocating multi-function resources in semiconductor wafer fabrication.